Chaotic time series forecasting based on Cdf9/7 biorthogonal wavelet kernel support vector machine

As biorthogonal wavelets have many advantages in signal processing, a new class of kernel function based on Cdf9/7 biorthogonal wavelet is proposed. The function has been proved to satisfy the admissible condition theoretically. Further, Cdf9/7 biorthogonal wavelet kernel support vector machine(SVM) is constructed to forecast the simulation data and stock market index with the character of chaos. The results of experiment show that compared with the general orthogonal wavelets kernel and non-orthogonal wavelets kernel, Cdf9/7 biorthogonal wavelet kernel SVM can not only avoid over-fitting effectively but also have higher forecasting accuracy and the ideal time performance.